LEGO-Eval/object_images
Overview This dataset contains 2D rendered images generated from 3D assets originally sourced from Objaverse and Objathor.The purpose of this dataset is to provide a large-scale collection of photo-realistic renderings for research on vision, multimodal learning, and text-to-3D understanding. Following prior works such as Diffusion4D and Stable-Zero123,we release only the rendered 2D images (not the original 3D assets) to facilitate efficient experimentation while preserving the… See the full description on the dataset page: https://huggingface.co/datasets/LEGO-Eval/object_images.
Overview
This dataset contains 2D rendered images generated from 3D assets originally sourced from [Objaverse](https://huggingface.co/datasets/allenai/objaverse) and [Objathor](https://github.com/allenai/objathor). The purpose of this dataset is to provide a large-scale collection of photo-realistic renderings for research on vision, multimodal learning, and text-to-3D understanding.
Following prior works such as [Diffusion4D](https://huggingface.co/datasets/hw-liang/Diffusion4D) and [Stable-Zero123](https://github.com/cvlab-columbia/zero123), we release only the rendered 2D images (not the original 3D assets) to facilitate efficient experimentation while preserving the provenance and licensing of the original Objaverse assets.
License and Attribution
Note: The renderings are generated with Objaverse. The renderings as a whole are released under the ODC-BY 1.0 license, consistent with the Objaverse dataset’s license terms. The licenses for the renderings of individual objects follow the same Creative Commons licenses as their corresponding source assets in Objaverse (e.g., CC-BY 4.0, CC-BY-NC 4.0, etc.).
Users must comply with these original licenses when redistributing or using the rendered images, including providing proper attribution to the original creators and sources.